Torsten Hoefler Peter Gottschling Wolfgang Rehm Andrew Lumsdaine Optimizing a Conjugate Gradient Solver with Non-Blocking Collective Operations. 374-382 2006 conf/pvm/2006 PVM/MPI http://dx.doi.org/10.1007/11846802_52 db/conf/pvm/pvm2006.html#HoeflerGRL06